Diagnosis of prostate cancer using cell-free DNA methylation profiles from expressed prostatic secretions.

389 Background: Urinary and seminal cell-free DNA (cfDNA) have been recognized as promising biomarkers of prostate cancer. DNA methylation signal in tumor is increasingly used as tumor diagnosis and longitudinal monitoring indicator. However, the clinical utility of cfDNA from expressed prostatic secretions (EPS) remains unknown. Methods: The prospective study includes 50 prostate cancer (PC) patients and equivalent benign prostatic hyperplasia (BPH) patients, where EPS samples were collected after the prostatic massage of each patient. Cell-free DNA from EPS was extracted and treated by PredicineEPIC, a liquid biopsy comprehensive DNA methylation assay. The study developed integrated bioinformatic algorithms to profile the genome-wide epigenomic characteristics and investigate tumor-specific methylation patterns. Results: The initial exploratory cohort included 5 PC and 11 BPH cases. PredicineEPIC whole-genome DNA methylation of EPS identified over 300 differentially methylated regions (DMRs). Prostate cancers were distinguished from the BPH group by an unsupervised clustering method, suggesting that the identified DMRs embody the specific profiles of malignant tumors. Additionally, this study detected genome-wide copy number variation burden (CNB) in parallel. A risk model was built for cancer risk assessment based on methylation and CNB characteristics. Conclusions: This study demonstrated the feasibility of methylation profiling of cfDNA in EPS in prostate cancer. This non-invasive liquid biopsy approach could sensitively navigate prostate cancer from benign prostatic hyperplasia, suggesting future direction of methylation-based liquid biopsy in detecting prostate cancer.